THEORETICAL AND DESCRIPTIVE LINGUISTICS

BY PETER THAMIN

Assignment Title

Critically evaluate the contribution of behaviourist psychology to our understanding of how humans learn and use language, paying particular attention to the Skinner-Chomsky debate.

Introduction

Language is learned, in the normal course of events, by children bright and dull, pampered or neglected.  It is recognised that a child who is capable of acquiring some particular human language is also capable of acquiring any language.  In Leonard Bloomfield’s words, “This is doubtless the greatest intellectual feat anyone of us is ever required to perform” (1933:29).  Appreciation of the enormity of this human capacity, given the intricacy and variety of languages of the world, has motivated an intense exploration of language learning by linguists and psychologists alike.  It is well understood by now that various theories of Language Acquisition may be set in the broader context of the empiricist-rationalist debate.  Thus, theories that assume the learner begins with almost no knowledge of what a language is, but has some simple kinds of “information-processing” mechanisms may be called an empiricist.  On the other hand, theory that assumes genetically determined specification of forms of language are rationalists theories.  This paper will only be focusing predominantly on the Skinner-Chomsky controversy.

Body

During the first half of the century, Skinner and other prominent behaviourists such as Russell, Thorndike, Bloomfield, Mowrer, Osgood etc, had put forward plausible views about Language Acquisition that became widely accepted in psychology during that period of time.  They held that one should study only external, measurable stimuli and responses and avoid talk of abstract and covert mental entities like “ideas” or “thoughts (Gardner, 1995: 32).  In other words, the behaviourists wanted to explain acquisition by assigning to the child very little innate behaviour – children only learn languages through processes such as imitation, probability, reinforcements, trial-and-error, associations, stimulus and responses (S-R principles) and the like.  According to Skinner (1957), no complicated innate mechanisms are needed in Language Acquisition.  All that is necessary is the systematic observation of the events in the external world which prompt the speaker to utter sounds.  This hypothesis was largely based on his empirical studies with rats and pigeons.  For example, in a typical experiment, a rat is put in a box containing a lever.  If it presses the lever, it is rewarded with a pellet of food.  Nothing forces the rat to press the lever initially, it happened only by chance or probability (linguistically, this would coincide with the “babble-luck” concept suggested by Thorndike 1943: 97, cited in Mowrer 1980: 98).  When the rat finds that food arrives, it presses the lever again.  Over time, the rat would have learnt that when it is hungry, it can obtain food by pressing the lever.  Then the task is made more difficult, the rat only gets rewarded if it presses the lever while a light is flashing.  At first, the rat is puzzled, but eventually, it learns the trick.  The task is then made more difficult again.

This type of learning is called operant conditioning by Skinner, or training by means of voluntary response.  Skinner suggests that it is by means of this mechanism that the vast majority of human learning takes place, including language learning:

“The basic processes and relations which give verbal behaviour its special characteristics are now fairly understood.  Much of the experimental work responsible for this advance has been carried out on other species, but the results have proved to be surprisingly free of species restrictions.  Recent work has shown that the methods can be extended to human behaviour without serious modification” (Skinner 1957: 3).

In other words, Skinner is proposing that all one needs to do in order to understand language is to identify the stimuli which will enable us to predict specific utterances (responses).

Furthermore, within the behaviourists’ school of thoughts, Russell claims that “The great majority of words are acquired by imitation, combined with the association between thing and word which the parent deliberately establish in the early stages” (1827: 50, cited in Mowrer 1980: 96).  For example, if you repeat a word, let’s say, “book” when you give a child a book enough times, then eventually the child will react to the word and associates its meaning accordingly through imitation and association (training or habit-formation as Chomsky would call it) since the object concerned “… does not exert gravitation. it does not nourish, it cannot bump onto the child’s head.  The effects which are shared by the word and the thing are those which depend upon the law of association or conditional reflexes or learned reactions” (ibid.).

An extended view taken by the behaviourists which is also based on the rats and pigeons experiment, would be one of reinforcement.  This basically suggests that children learn from their caretakers (most commonly mothers) since they provide not only the obvious primary reinforcers such as food, contact and other comforts, but because of their association with those primary reinforcers, how their very presence and other behaviours, including speech and other sound production, become conditioned reinforcers.  This means that the mother is in a position to reinforce a good deal of verbal behaviour on the part of the infant.

According to Chomsky and his associates, all of the examples given so far by behaviourists is patently irrelevant to human language and most importantly, they have fundamentally misunderstood the complexity and dynamicity of the nature of language.  For example, in Chomsky’s extensive review of B.F Skinner’s “Verbal Behaviour”, he began by stating the simple and well-defined sequence of events observed in the boxes of rats is just not applicable to language.  The terminologies used also cannot be applied to human languages without becoming ambiguous and hopelessly vague.

For example, consider first Skinner’s use of the notion of S-R principles.  A typical example of “stimulus control” for Skinner would be that a piece of music or a painting (stimulus) would create utterances such as Mozart or Dutch (response).  However, according to Chomsky, these hypotheses “covers almost no aspect of linguistic behaviour” (Chomsky 1959: 31).  How do we know that these are the predictive responses?  We might say instead, It doesn’t tickle my fancy, I like R&B, Remember our camping trip last summer etc.  Nevertheless, Skinner would claim that each of these responses is under the control of some other stimulus.  In other words, if we look at the “red chair” and uttered the word red, the response is under the control of the stimulus “redness”.  Again, Chomsky argues that if this was the case, then the word “stimulus” would have lost all of its objectivity in this usage since we cannot predict verbal behaviour in terms of the stimuli in the speaker’s environment anymore because we do not know what the current stimuli are until it responds (p. 32).  Chomsky holds Skinner severely accountable for hypothesising the S-R principles because “this device is as simple as it is empty.  Since properties are free for the asking” (p. 31).

Secondly, the traditional and highly accepted view of the theory of imitation in Language Acquisition has also been critically rebutted.  For example, how do we account for the fact that children tend to overgeneralised the verb goed instead of went, rationalising the past tense of verbs is generally formed by adding an ed ending?  And since adults do not use this form, how can we say that children acquire language through imitation or habit-formation or association etc?  Also, as the well-known Jean Berko (1958, cited in Cofer and Musgrave 1963) experiment have evidently proved, the fact that children on one had would automatically say the voiceless sibilant /-s/ as in cats and lips if a plural word ends with a voiceless consonant, while on the other, they would say the voiced sibilant /-z/ as in dogs and ribs, if a word ends with a voiced consonant, indicates clearly they are not imitating.

In regards to reinforcement, it is not surprising to see that Chomsky inevitably finds skinner’s functional definition of reinforcement or reward (since a reward reinforces a behaviour being learnt) unsatisfactory, firstly because, according to Skinner, a child who talks to himself (which happens quite frequently) can automatically reinforce themselves because they know they are producing sounds which they have heard before in the speech of others and secondly, because a poet who is uttering words aloud can also be reinforced by the knowledge that others will be influenced by the poerty in the future.  From these two examples, Chomsky rightly claims that “the notion of reinforcement has totally lost whatever meaning it may ever have” (Chomsky 1959: 39).  That is, the notion that “X” is reinforced by “Y” is being extensively used to mean that the reinforcing “stimulus” need not impinge on the “reinforced person” or need not even exist (it is sufficient that it be imagined or hope for).  Even if we were to use the term reinforcement appropriately, it is quite clear that children do not receive pellets of food when making correct utterances (Aitchisn 1989: 9).  Furthermore, if we were to believe that “slow and careful” reinforcement applied by their mothers which “meticulous care” is necessary for Language Acquisition as suggested by Skinner, then how do we explain the phenomenon that children continually produce words and sentences that they have never heard before?  If a child acquires language through reinforcement, we could never account for the rate if acquisition that occurs (Chomsky 1959: 42).

Evidently, according to the nativists, the empiricists view that Language Acquisition can be explained by empirical data or more precisely, by observing the behaviour of experimental rats and pigeons is plainly inadequate.  Therefore, Chomsky and his colleagues properly claimed that a more plausible theory of Language Acquisition must address the fundamental phenomenon that children acquire an entire grammatical system which enables them to produce and understand a potentially infinite number of utterances in their first few years of life.  Thus, Chomsky proposes that the nature of language is – (i) rule-governed and creative; and (ii) the rule-governed system is stored in the “Language Acquisition Device” (LAD – Chomsky 1957, 1965 which later was called “Universal Grammar”).  In other words, this is innate.

As we can see, Chomsky expresses a completely different position.  In his view, he believes language is an extremely rich and complex system which is more than a series of associations between words that are linearly ordered.

Let us then consider, for example, the sentence: “Is the boy who left happy?”.  The sentence consists of a moved auxiliary “is” that agrees in number with Noun Phrase (NP) “the boy”, which in turn is modified by the relative clause “who left”.  The sentence consists of a moved auxiliary “is” that agrees in number with the Noun Phrase (NP) “the boy”, which in turn is modified by the relative clause “who left”.  Now, there is nothing in the phonetic string of the sentence that tells the child about the structure or the syntactic rules of creating the interrogative [Is the boy who left happy?] (Ingram, 1989: 26).  This, in fact, can be seen as an important observation which diminishes the behaviourist’s theory that language is acquired through external exposure.  Consider another example.  Suppose a child noted that in English, through imitation and experience, that one can convert a declarative into an interrogative by simply moving the verb from a middle position into a front position.  Therefore, “The man is here” would become “Is the man here?”.  If this was the case, the declarative “The man who is here is tall” would then be converted into the ungrammatical interrogative *”Is the man who here is tall?”.  However, according to Chomsky, children do not utter these ill-formed sentences, instead they would naturally generate the correct utterances “Is the man who is here tall?” simply because language is so programmed in the human mind, the implicit knowledge that “the man who is here” is a single, indissoluble NP, is a feature of every child’s linguistic competence which could not conceivably have been induced from experience.  Therefore, this kind of knowledge about the possible structure of the language, according to Chomsky, must be rule-governed and it must be stored in the child’s LAD.

What is an LAD?  An LAD is an abstract mechanism which receives certain corpus of utterance.  Many of these utterances are grammatical sentences in the language to which LAD is exposed.  Given such a corpus, LAD is so constructed that it can develop a theory of the regularities that underlie the speech to which it has been exposed.  However, McNeil (1970: 150) suggests that whatever the LAD contains, it must be universally applicable, for LAD must be able to acquire any language, hence, the term “Universal Grammar”.  But Universal Grammar is not a set of properties that are true of all languages.  Rather, it is a set of specifications for the shape of permissible kinds of languages.  In other words, at a certain stage, the child’s grammar will correspond to a possible adult grammar which he will ultimately master.  The following example borrowed from Aitchison (1989: 28) ultimately illustrates this point clearly:

(a) I gave a book to Mary

(b) I gave Mary a book

(c) I reported the crime to the police

(d) * I reported the police the crime

By examining the sentences above, we can propose a theory that suggests there is a rule of English called Dative which moves an indirect object next to the verb as in (b).  Dative has not been applied in (a) and (c).  And we also need to restrict the Dative rule so that it does not apply to (d).  To do this, we can mark the verb “report” as an exception to our Dative.

Now consider this, how does the child ever learn that (d) is ungrammatical?  The child cannot learn by imitation since he will only hear sentences like (a), (b) and (c).  Thus, this strongly suggests that language is a rule-governed system and it is innate in every child’s mind/brain:

Given such facts, it is natural to postulate that the idea of a rule-governed system is part of the innate schematism applied by the mind to the data of experience (Chomsky 1972a: 30).

Moreover, creativity is also another fundamental aspect of language which Chomsky repeatedly emphasised.  Take, for example, the sentence The people who called and wanted to rent your house when you go away next year are from California (Miller and Chomsky, 1963, cited in McNeil 1968).  If we assume that a person gradually accumulates string of words throughout their life and stores them ready for use on appropriate occasions, then, this means that the child would need to go deep into his or her brain and choose the right words at the right sequence which he or she has heard before, every time that he or she wishes to speak.  This assumption would simply be illogical and absurd!  The difficulties with learning through imitations are simple and arithmetical (refer to McNeil 1968 for substantial proof).  As a result, Chomsky made this plausible conclusion:

Having to master a language, one is able to understand an infinite number of expressions that are new to one’s experience, that bear no simple physical resemblance and are in no simple way analogous to the expressions that constitute one’s linguistic experience… The normal use of language is, in this sense, a creative activity.  This creative aspect of normal language is one fundamental factor that distinguishes human language from any known system of animal communication (Chomsky 1972: 100).

Finally, Chomsky also believes that language is not simply the association of words and meaning.  Instead, he believes that language consists of two levels of representation –